Actually pharma companies have used computer simulations to create new drugs from existing ones for a while, but the search was guided mostly by the theory of pharmacophores, which was a more or less empirical theory. What's really cool about Alán's work is that he's replaced the pharmacophore theory with a deep neural network, so it can actually "learn" what works and what doesn't and come up with new answers that would be unavailable otherwise.
Yeah, it's a really exciting development. Hopefully soon the big pay off is new and effective treatments for drug resistant strains of viruses and bacteria.
Actually pharma companies have used computer simulations to create new drugs from existing ones for a while, but the search was guided mostly by the theory of pharmacophores, which was a more or less empirical theory. What's really cool about Alán's work is that he's replaced the pharmacophore theory with a deep neural network, so it can actually "learn" what works and what doesn't and come up with new answers that would be unavailable otherwise.
Yeah, it's a really exciting development. Hopefully soon the big pay off is new and effective treatments for drug resistant strains of viruses and bacteria.